Philips Respironics System One REMstar Pro file format

BryanC wrote:

builta wrote:Hi,

I tried uploading and looking at the output about a month ago but, unfortunately at the time, it did not provide a meaningful output for me. you see, I am using a "PR System One BiPAP Auto" and the routines were set up for an APAP.

My question is, has the web site been updated to deal with BiPAP Auto or is it still only working for CPAP and CPAP Auto?


Cheers

ohh, that i couldn't tell ya, you'll have to wait for cpapdork to reply.... BUT, is there any way you could email me a bunch of your files? I have the PRS1 CPAP Flex+ Pro (I believe? only had it about a month, but I know it's not a BiPap)... I'm sure yours has a lot of data I haven't got 'details' for yet, and it would help me figure out how to support the BiPap machines too...
If you can, I need the 001, 002, and 005 files (and, if there are others? (003, 004, 007, etc.) for each? (at this point, i don't know!))
bac512@users.sourceforge.net

Hi,

A month's worth of files on it's way.

Cheers

Hi cpapdork/bryanc

i've got over a years worth of data on my SD , but uploading with a max of 50 files at a time is a bit of an effort ...

Also if it helps , i have a couple of weeks data , and the associated Encore Pro report that matches , as processed by my 'specialist' if that may help identify some of those 'unknown events' PM me an email and i can send them through...

Regards GarFin.

garfin wrote:Hi cpapdork/bryanc

i've got over a years worth of data on my SD , but uploading with a max of 50 files at a time is a bit of an effort ...

Also if it helps , i have a couple of weeks data , and the associated Encore Pro report that matches , as processed by my 'specialist' if that may help identify some of those 'unknown events' PM me an email and i can send them through...

Regards GarFin.

Hi. You can use a program like winzip to condense all the files into one and hen send just the zipped file. That's what I did.

Cheers

builta wrote: Hi. You can use a program like winzip to condense tall the files into one and hen send just the zipped file. That's what I did.

Cheers

Thanks, but i was referring to uploading into cpapdork's project.. http://www.hoolehan.com/onkor/

BryanC, you should have my Encore report and dataset in yr email by now ... I have a futher 1year+ of data if you want it , just no corresponding Encore Report.

builta wrote:My question is, has the web site been updated to deal with BiPAP Auto or is it still only working for CPAP and CPAP Auto?

No update for supporing BiPAP machines. I don't have a BiPAP machine so I can't get at the data files. If you're willing, you can send me a batch of files and I will post on the Onkor wiki. If one of us has some time, we can take a look to see if the formats are similar and if they can be supported in onkor.

garfin wrote: weeks data , and the associated Encore Pro report that matches , as processed by my 'specialist' if that may help identify some of those 'unknown events' PM me an email and i can send them through...

Yes! Having the data and some known reports of that data is particularly valuable. It lets us "check" our results.

I think some of you are sending data to me and/or BryanC. Just an FYI, we're not specifically working together (at least not as far as I know). However, BryanC, maybe even if we work separately, we could agree to post "sanitized" datasets on the Onkor wiki so that anyone (and everyone) can benefit from these data. Or we could put it on an independent (non-onkor) site if you like as well.

We've all got the same goal I think, so it would be great to share info as much as possible.

cpapdork wrote:

garfin wrote: weeks data , and the associated Encore Pro report that matches , as processed by my 'specialist' if that may help identify some of those 'unknown events' PM me an email and i can send them through...

Yes! Having the data and some known reports of that data is particularly valuable. It lets us "check" our results.

I think some of you are sending data to me and/or BryanC. Just an FYI, we're not specifically working together (at least not as far as I know). However, BryanC, maybe even if we work separately, we could agree to post "sanitized" datasets on the Onkor wiki so that anyone (and everyone) can benefit from these data. Or we could put it on an independent (non-onkor) site if you like as well.

We've all got the same goal I think, so it would be great to share info as much as possible.

yeah, absolutely. we've all got the same goal, and the more we all know, the quicker we'll be able to figure more out.... and since I have no site, using the onkor site is fine by me.

I donno if you've looked yet or not at the properties file, but in it I see BASICKEY, DETAILSKEY, ERRORKEY, and VALIDCHECK. I don't really know what they'd be used for, since all the data appears to be unencrypted, but maybe they're used for checksums?

I've noticed(or so it appears with my not-so-advanced processing of the files) I find a lot with errors in them. Like one I was looking at yesterday, all the data was there, but one of the records had all 0's for the header.... If you're removing these from your datasets, you should at least keep them in a separate directory, for error checking later.... (I found this because my program reads the header, then the data, then adds the record length to the start position to find the next record... and with it all 0's in the header... it just kept reloading that record! )
and FYI, in file 002, donno if you've gotten the Bipap files yet, but it seems 'event' 3 is

0x03 BIPRESSURE [16bit TIME], [8bit PRESSURE_LOW1],[8bit PRESSURE_HIGH1]

Thanks everyone, especially Sleepy, for your help with leak info. I've updated the onkor website (http://www.hoolehan.com/onkor/) to include a plot of Leak (in liters per minute) as well as annotations for Large Leaks.

Hi..

Question regarding Leaks.. Has anyone figured out how to actually interpret that byte?

When I subtract 17/18 from it, my leak graphs visually match Encore, and my leak averages mostly adds up. Though Something doesn't feel right about doing that. I've tried subtracting the average to get a neater plot. Looks good, but then the average doesn't look right. Divide by ten doesn't work either.

Can anyone clue me in it?

As for CSR, the first of the CSR data fields is definitely 16bit. both data values subtract from current time.. I have a visual & time match with Encore in the sleep flags plot and the waveform overlay.

Regarding my theory on the 16bit time delta.. It doesn't really matter which way it's handled, as it's capped at an 8bit value of 2 minutes. I've always felt "claustrophobic" using 8bit time deltas in embedded projects, hence why I'm so anal about it.

Apart from leaks, the only other thing I can't figure out is why the VSNORE indice value is so high.. Mine doesn't add up, even though the sleep flags match encore perfectly (using combined 0x0D & (0x11 data field2>0) vsnore events) Perhaps Encore uses this >0 vsnore value as a time duration, and I calculate this like % of night in CSR?

By the way, 0x0D VSNORE events are the only ones that change pressure.

jedimark wrote:Hi..

Question regarding Leaks.. Has anyone figured out how to actually interpret that byte?

When I subtract 17/18 from it, my leak graphs visually match Encore, and my leak averages mostly adds up. Though Something doesn't feel right about doing that. I've tried subtracting the average to get a neater plot. Looks good, but then the average doesn't look right. Divide by ten doesn't work either.

Can anyone clue me in it?

As for CSR, the first of the CSR data fields is definitely 16bit. both data values subtract from current time.. I have a visual & time match with Encore in the sleep flags plot and the waveform overlay.

Regarding my theory on the 16bit time delta.. It doesn't really matter which way it's handled, as it's capped at an 8bit value of 2 minutes. I've always felt "claustrophobic" using 8bit time deltas in embedded projects, hence why I'm so anal about it.

Apart from leaks, the only other thing I can't figure out is why the VSNORE indice value is so high.. Mine doesn't add up, even though the sleep flags match encore perfectly (using combined 0x0D & (0x11 data field2>0) vsnore events) Perhaps Encore uses this >0 vsnore value as a time duration, and I calculate this like % of night in CSR?

By the way, 0x0D VSNORE events are the only ones that change pressure.

from what I've seen, my leak graph matches encore's exactly, with no modifications.... I mean, apart from my graphs not being the same length as encore's...

as for the VSNORE, how are your %'s coming out? maybe encore only adds it against the normal breathing time- all events? (or + all enets?) maybe only the 0x0D events count towards the total?

as for CSR, I can see where that could be longer than the 120 seconds, much longer, requiring 2 bytes, but have you seen it longer? or does it report it too every 2 minutes?
I GUESS it's possible, if it only reports that items at the END of their run that they (OA, CA, CSR, VS, etc.) could need 2 bytes....

jedimark wrote:

When I subtract 17/18 from it, my leak graphs visually match Encore, and my leak averages mostly adds up.

Do you have Encore set to show total leak? Or is it taking off something as a base vent rate first?

I ask because the graphs and numbers match for me, and mine is set for total leak. The only difference was a .3 rounding on Average Leak, but I think that's because Respironics throws out about six minutes on the machine and in the software (or seems to...) so datasets are more complete in the 002 file.

BryanC wrote:

as for CSR, I can see where that could be longer than the 120 seconds, much longer, requiring 2 bytes, but have you seen it longer?

I have had periodic breathing (the CSR type of waxing and waning) for longer than 2 minutes, I'll look and see one of those nights and see how it gets recorded.

-- SleepyCPAP

Hi,

In trying some experiments on PRS1 xPAPs I found something that you may need to be aware of. If a blank SD card (or one that has been wiped after review- ref our kindly DMEs) or a card from a different PRS1 machine is inserted the PRS1 xPAPs will load a condensed data file set that contains all the daily data (minus wave form data) in a single file set.

In the Machine ID folder you will get three folders and files of potential interest:

1. In the directory you will find a binary "sync" file like "222e1d79-732a-46ff-a049-ad2d8eb54d47.sync"
2. In directory p0 you will get a "0000000001.001" and a "0000000xxx.002" where xxx = a number equal to the number of data sets recorded?
3. In directory e your will get a "000.004.uploaded"

These files when mixed with other normal files could cause confusion unless you understand and treat them separately and as condensed multi-day files.

Cheers

As for any leak settings in EncorePro, I haven't seen them. I'll have a look through the menus. My minimum leak value=15, the maximum=48, and weighted average=20.15. Yet encore reports my leaks as 0.7 - there must be some baseline calc I am missing.

I've noticed Encore uses a rounding function that rounds up when >=.4, not >=.5 like most people expect.

Encore reports my VSNORE index around 10.3, mine calculations come out to 3.89.. I have just been taking the count of vsnore events in a time period and dividing by TotalTime in hours. Even when adding both types together it doesn't work out to 10.3 - once again I'm missing something.

I was wrong about 0x0D VSNORE events being included in Encores sleepflags.. But they are the VSNORE event shown in Encore's waveform overlay.
My CSR events start and end at exactly the same time as shown in Encores WaveFormReport. I have noticed sometimes CSR areas can overlap.

Well, found the leak setting in Encore.. It shows two methods.. so my initial calcs were right. Wish I knew that sooner.

I have a suspicion the other method Encore uses for leak display subtracts the 7 day leak average value from each sample (and crops below zero)
There is some good clues in the Encores "calculations" screen.

Still haven't figured out the vsnore calc thing..

I'm having issues with the events and breathing patterns. seems no matter how I adjust them, (specifically dealing with Central Apneas in this case) there doesn't seem to be a standard way to add or subtract the CA event's value, it doesn't consistently match up with the beginning or end of the event. sometimes it'll even completely miss the 'apparent' apnea. and, I would think it was just scaling it wrong maybe, but one will match, then next won't, maybe a few, then another will, etc.... has anyone come up with some method that seems to work ALL the time?